The 3-0 issue always seemed to me like it could use some further refinement and analysis, beyond just the suggestion that it's a once every 33 series phenomenon.
It seems to me there are a spectrum of situations where teams are down 3-0. Often (usually), a team is down 3-0 because they are significantly inferior. In those situations, the likelihood of coming back is very low.
But sometimes, even between two evenly matched teams, one of them can get down 3-0. In fact, in a series between two evenly matched teams, you would expect one or the other of the teams to lose the first three about 25 percent of the time. Indeed, it can even happen in a non-negligible number of cases that a slightly better team will lose the first three.
I think the 2004 ALCS was this latter kind of case. That is, the 25 percent of the time, or so, where two teams are evenly matched but one just happens to have won the first three games. Now, it's true, that team continues to have a significant advantage. But it's not quite as bad as the 1/33 suggests. The chance of a team evenly matched with its opponent winning four in a row is about 6.25 percent -- so, in this latter case it should happen about once every 16 times.
I especially hate when a team down 3-0 wins game 4 and you continue to hear about how few teams have ever come back from 3-0 down, as though it still matters.
Anyway, when I started typing this drivel, my point was going to be that I thought this series fell in that latter category -- the category where two close to evenly matched teams has just happened to win 3 in a row, making this a series where the scary 1/33 stat doesn't seem as daunting. As I've been typing, however, it seems clear that I have neglected to take into account the fact that team down 3-0 is susceptible to severe sphincter tightening, like Orioles, who appear to be playing very shitty.